Forward Collision Avoidance Maneuver Selection for Multiple Obstacles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current forward collision avoidance systems in vehicles only evaluate collision risks with one obstacle at a time, failing to consider how maneuvers to avoid one obstacle might result in collisions with others, leading to underestimated threats in multiple obstacle scenarios and computationally demanding algorithms.
Innovation Solution
A method that establishes the presence and position, velocity, and acceleration of obstacles, evaluates maneuvers that lead to collisions with each obstacle, forms the union of these maneuvers, and determines the set of maneuvers that can avoid collisions, enabling autonomous decision-making and execution of collision avoidance maneuvers in multiple obstacle scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the system evaluates collision risks with only one obstacle at a time, then the decision algorithm is computationally simple, but the collision threat is underestimated in multiple obstacle scenarios
Solution Approach 1:
The patent segments the evaluation process into two distinct phases: (1) individual obstacle assessment where each obstacle is evaluated separately to determine collision risk, and (2) integrated decision-making where all individual assessments are combined to form a comprehensive collision threat evaluation. This segmentation allows the system to maintain computational simplicity in each phase while achieving reliable overall assessment through the combination of all obstacle evaluations.
2Reliability
If the system considers all possible scenarios with multiple obstacles, then the collision threat evaluation is comprehensive, but the computational demand increases exponentially
Solution Approach 1:
The patent divides the complex multi-obstacle evaluation into separate single-obstacle assessments. Each obstacle is evaluated independently to determine if it poses a collision risk, and these individual evaluations are then integrated in the decision algorithm. This segmentation avoids the exponential computational complexity of evaluating all possible multi-obstacle scenarios while maintaining comprehensive threat assessment through the aggregation of individual obstacle risks.
3Device complexity
If the system evaluates each obstacle separately without considering interactions, then the computational complexity is reduced, but maneuvers to avoid one obstacle may result in collisions with others
Solution Approach 1:
The patent combines individual obstacle assessments with integrated decision-making. The decision algorithm merges all separate obstacle evaluations to determine the overall collision threat and select avoidance maneuvers that consider the positions and risks of all obstacles simultaneously. This merging ensures that maneuvers are evaluated for their impact on multiple obstacles, preventing situations where avoiding one obstacle causes collision with another.
Data Source
AI summary
A method and system for autonomous collision avoidance in multiple obstacle scenarios. A forward collision avoidance system detects obstacles in front of a vehicle hosting the system and estimates a position, a velocity and an acceleration of each of the obstacles. The maneuvers which the vehicle is capable of performing that will lead to a collision with each respective obstacle are evaluated separately. The union of maneuvers which will lead to collision with any one of the obstacles is formed. The set of maneuvers which the vehicle is capable of performing through which collision with any of the obstacles may be avoided is established, and the set used for deciding how to avoid or mitigate collision with any one of the obstacles. Finally, a collision avoidance maneuver based on the decision is executed autonomously.